Halfway to the Peak: The JWST MIRI 5.6 micron number counts and source population

TL;DR

This study analyzes JWST MIRI 5.6 micron images, revealing detailed galaxy counts, redshift distributions, and the discovery of new dusty galaxy populations at high redshift, demonstrating MIRI's capability to probe cosmic noon and beyond.

Contribution

First comprehensive analysis of JWST MIRI 5.6 micron counts, linking them to galaxy populations and models, and identifying new dusty high-redshift galaxies.

Findings

Number counts show a pronounced knee at ~2 uJy.

Most MIRI sources are moderate to low mass star-forming galaxies at z~0.5-2.

Discovered 31 new MIRI-bright sources, many likely dusty galaxies at z>3.

Abstract

We present an analysis of eight JWST Mid-Infrared Instrument (MIRI) 5.6 micron images with $5\,\sigma$ depths of ~0.1 uJy. We detect 2854 sources within our combined area of 18.4 square arcminutes. We compute the MIRI 5.6um number counts including an analysis of the field-to-field variation. Compared to earlier published MIRI 5.6 um counts, our counts have a more pronounced knee, at roughly 2 uJy. The location and amplitude of the counts at the knee are consistent with the Cowley et al. (2018) model predictions, although these models tend to overpredict the counts below the knee. In areas of overlap, 84% of the MIRI sources have a counterpart in the COSMOS2020 catalog. These MIRI sources have redshifts that are mostly in the $z\sim0.5-2$, with a tail out to $z\sim5$. They are predominantly moderate to low stellar masses ($10^8-10^{10}$M$_{\odot}$) main sequence star-forming galaxies, suggesting that with ~2hr exposures, MIRI can reach well below $M^*$ at cosmic noon and reach higher mass systems out to $z\sim5$. Nearly 70% of the COSMOS2020 sources in areas of overlap now have a data point at 5.6um (rest-frame near-IR at cosmic noon) which allows for more accurate stellar population parameter estimates. Finally, we discover 31 MIRI-bright sources not present in COSMOS2020. A cross-match with IRAC channel 1 suggests that 10-20% of these are likely lower mass (M$_*\approx10^9$M$_{\odot}$), $z\sim1$ dusty galaxies. The rest (80--90%) are consistent with more massive but still very dusty galaxies at $z>3$.